An Approach Towards Image Edge Detection Based on Interval-Valued Fuzzy Mathematical Morphology and Admissible Orders

Author(s)

Sussner, Peter ; Carazas, Lisbeth Corbacho ; Novak, V ; Marik, V ; Stepnicka, M ; Navara, M ; Hurtik, P

• Document type: Journal article
• Source: PROCEEDINGS OF THE 11TH CONFERENCE OF THE EUROPEAN SOCIETY FOR FUZZY LOGIC AND TECHNOLOGY (EUSFLAT 2019); v. 1, p. 8-pg., 2019-01-01.

Abstract

Edge detection is an important step for preprocessing digital images before more advanced methods of image analysis such as segmentation can be applied. There are an infinite number of edge detectors that can be derived from pairs of fuzzy dilation and erosion operators. Usually, an edge detector is based on the incorrect assumption that there is no uncertainty regarding the pixel values of the given digital image. The approaches presented in this paper do not rely on this assumption. Instead, the uncertainty regarding the pixel values is modelled in terms of an interval-valued image. After an application of an interval-valued fuzzy dilation and erosion, we are able to produce a binary edge image after a number of steps including an order-preserving transformation based on an admissible order. (AU)

• FAPESP's process: 18/13657-1 - Some lattice computing approaches towards computational intelligence, image processing and analysis
• Grantee: Peter Sussner
• Support Opportunities: Regular Research Grants